Some of the most important players in the artificial intelligence industry have created a consortium with the aim of ending NVIDIA’s dominance. The names involved are absolutely prominent: the list includes Samsung, Google, Qualcomm, Intel and Arm. The consortium is called the Unified Acceleration, or UXL, Foundation, and is working on an open-source machine learning software suite free of any kind of NVIDIA technology. The consortium itself expects to reach the maturity of the software by the second half of the year, although a launch window has not yet been specified.
According to the official information that has emerged so far, the project includes OneAPI, a standard created by Intel that does not require the use of specific programming languages and related architectures, such as NVIDIA CUDA (Compute Unified Device Architecture), the parallel processing system through which the GPU cores can be exploited to accelerate operations different from those originally planned.
If in the short term the UXL Foundation wants to offer developers an “NVIDIA FREE” alternative to AI operations, in the long term it hopes to be able to incorporate the hardware and software of the company led by Jensen Huang into its platform as well. The consortium says it’s also trying to engage Amazon and Microsoft to make sure its solution is compatible with every type of chip.
Microsoft, one of the companies that has thrown itself into the AI sector with the most enthusiasm (impossible not to mention the very rich partnership with OpenAI, from which Copilot was born) is a very, very illustrious absence. And AMD is also missing, which is the “other” big player in the discrete GPU sector. It’s worth mentioning that it was rumored last year that these very two companies had partnered with more or less the same goal, but more focused on hardware – in other words, to make alternative AI chips to NVIDIA’s.